A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence

A hybrid scheme that utilizes MPI for distributed memory parallelism and OpenMP for shared memory parallelism is presented. The work is motivated by the desire to achieve exceptionally high Reynolds numbers in pseudospectral computations of fluid turbulence on emerging petascale, high core-count, ma...

Descripción completa

Detalles Bibliográficos
Autores principales: Mininni, P.D., Rosenberg, D., Reddy, R., Pouquet, A.
Formato: JOUR
Materias:
Computational fluids
MPI
Numerical simulation
OpenMP
Parallel scalability
Central component
Code performance
Computational fluid
Discretizations
Distributed Memory
Domain decompositions
Fluid turbulence
Global data
High Reynolds number
Hybrid implementation
Hybrid paradigm
Hybrid scheme
Massively parallel processing systems
Maximum Efficiency
Optimal number
Performance tests
Petascale
Pseudospectral
Shared memory parallelism
Application programming interfaces (API)
Domain decomposition methods
Fast Fourier transforms
Fluids
Optimal systems
Reynolds number
Scalability
Turbulence
Parallel processing systems
Acceso en línea:http://hdl.handle.net/20.500.12110/paper_01678191_v37_n6-7_p316_Mininni
Aporte de:
Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA) de Universidad de Buenos Aires
id todo:paper_01678191_v37_n6-7_p316_Mininni
record_format dspace
spelling todo:paper_01678191_v37_n6-7_p316_Mininni2023-10-03T15:05:24Z A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence Mininni, P.D. Rosenberg, D. Reddy, R. Pouquet, A. Computational fluids MPI Numerical simulation OpenMP Parallel scalability Central component Code performance Computational fluid Discretizations Distributed Memory Domain decompositions Fluid turbulence Global data High Reynolds number Hybrid implementation Hybrid paradigm Hybrid scheme Massively parallel processing systems Maximum Efficiency MPI OpenMP Optimal number Parallel scalability Performance tests Petascale Pseudospectral Shared memory parallelism Application programming interfaces (API) Domain decomposition methods Fast Fourier transforms Fluids Optimal systems Reynolds number Scalability Turbulence Parallel processing systems A hybrid scheme that utilizes MPI for distributed memory parallelism and OpenMP for shared memory parallelism is presented. The work is motivated by the desire to achieve exceptionally high Reynolds numbers in pseudospectral computations of fluid turbulence on emerging petascale, high core-count, massively parallel processing systems. The hybrid implementation derives from and augments a well-tested scalable MPI-parallelized pseudospectral code. The hybrid paradigm leads to a new picture for the domain decomposition of the pseudospectral grids, which is helpful in understanding, among other things, the 3D transpose of the global data that is necessary for the parallel fast Fourier transforms that are the central component of the numerical discretizations. Details of the hybrid implementation are provided, and performance tests illustrate the utility of the method. It is shown that the hybrid scheme achieves good scalability up to ∼20,000 compute cores with a maximum efficiency of 89%, and a mean of 79%. Data are presented that help guide the choice of the optimal number of MPI tasks and OpenMP threads in order to maximize code performance on two different platforms. © 2011 Elsevier B.V. All rights reserved. Fil:Mininni, P.D. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales; Argentina. JOUR info:eu-repo/semantics/openAccess http://creativecommons.org/licenses/by/2.5/ar http://hdl.handle.net/20.500.12110/paper_01678191_v37_n6-7_p316_Mininni
institution Universidad de Buenos Aires
institution_str I-28
repository_str R-134
collection Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA)
topic Computational fluids
MPI
Numerical simulation
OpenMP
Parallel scalability
Central component
Code performance
Computational fluid
Discretizations
Distributed Memory
Domain decompositions
Fluid turbulence
Global data
High Reynolds number
Hybrid implementation
Hybrid paradigm
Hybrid scheme
Massively parallel processing systems
Maximum Efficiency
MPI
OpenMP
Optimal number
Parallel scalability
Performance tests
Petascale
Pseudospectral
Shared memory parallelism
Application programming interfaces (API)
Domain decomposition methods
Fast Fourier transforms
Fluids
Optimal systems
Reynolds number
Scalability
Turbulence
Parallel processing systems
spellingShingle Computational fluids
MPI
Numerical simulation
OpenMP
Parallel scalability
Central component
Code performance
Computational fluid
Discretizations
Distributed Memory
Domain decompositions
Fluid turbulence
Global data
High Reynolds number
Hybrid implementation
Hybrid paradigm
Hybrid scheme
Massively parallel processing systems
Maximum Efficiency
MPI
OpenMP
Optimal number
Parallel scalability
Performance tests
Petascale
Pseudospectral
Shared memory parallelism
Application programming interfaces (API)
Domain decomposition methods
Fast Fourier transforms
Fluids
Optimal systems
Reynolds number
Scalability
Turbulence
Parallel processing systems
Mininni, P.D.
Rosenberg, D.
Reddy, R.
Pouquet, A.
A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence
topic_facet Computational fluids
MPI
Numerical simulation
OpenMP
Parallel scalability
Central component
Code performance
Computational fluid
Discretizations
Distributed Memory
Domain decompositions
Fluid turbulence
Global data
High Reynolds number
Hybrid implementation
Hybrid paradigm
Hybrid scheme
Massively parallel processing systems
Maximum Efficiency
MPI
OpenMP
Optimal number
Parallel scalability
Performance tests
Petascale
Pseudospectral
Shared memory parallelism
Application programming interfaces (API)
Domain decomposition methods
Fast Fourier transforms
Fluids
Optimal systems
Reynolds number
Scalability
Turbulence
Parallel processing systems
description A hybrid scheme that utilizes MPI for distributed memory parallelism and OpenMP for shared memory parallelism is presented. The work is motivated by the desire to achieve exceptionally high Reynolds numbers in pseudospectral computations of fluid turbulence on emerging petascale, high core-count, massively parallel processing systems. The hybrid implementation derives from and augments a well-tested scalable MPI-parallelized pseudospectral code. The hybrid paradigm leads to a new picture for the domain decomposition of the pseudospectral grids, which is helpful in understanding, among other things, the 3D transpose of the global data that is necessary for the parallel fast Fourier transforms that are the central component of the numerical discretizations. Details of the hybrid implementation are provided, and performance tests illustrate the utility of the method. It is shown that the hybrid scheme achieves good scalability up to ∼20,000 compute cores with a maximum efficiency of 89%, and a mean of 79%. Data are presented that help guide the choice of the optimal number of MPI tasks and OpenMP threads in order to maximize code performance on two different platforms. © 2011 Elsevier B.V. All rights reserved.
format JOUR
author Mininni, P.D.
Rosenberg, D.
Reddy, R.
Pouquet, A.
author_facet Mininni, P.D.
Rosenberg, D.
Reddy, R.
Pouquet, A.
author_sort Mininni, P.D.
title A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence
title_short A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence
title_full A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence
title_fullStr A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence
title_full_unstemmed A hybrid MPI-OpenMP scheme for scalable parallel pseudospectral computations for fluid turbulence
title_sort hybrid mpi-openmp scheme for scalable parallel pseudospectral computations for fluid turbulence
url http://hdl.handle.net/20.500.12110/paper_01678191_v37_n6-7_p316_Mininni
work_keys_str_mv AT mininnipd ahybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
AT rosenbergd ahybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
AT reddyr ahybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
AT pouqueta ahybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
AT mininnipd hybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
AT rosenbergd hybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
AT reddyr hybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
AT pouqueta hybridmpiopenmpschemeforscalableparallelpseudospectralcomputationsforfluidturbulence
_version_ 1807321733406916608

Ejemplares similares